when viewed historically. However, it is also clear that the continued high fertility in those countries has represented a costly lost opportunity for raising levels of living at rates consistent with remarkably effective productive performance. If the less developed regions could raise their current per capita income growth rates by one third, it would reduce their per capita income doubling time from somewhat over 25 years to 18 years. Under current circumstances this could be accomplished entirely through a fall of the average birth rate in the less developed regions from their roughly 40 per 1,000 level to 30 per 1,000, a 25 percent shift, which, in addition to its income effects, could have perhaps equally large family-welfare effects not captured by conventional income measures.
A larger decline in birth rates, down to the level of the developed regions—less than 20 per 1,000—would raise the stakes commensurately, again quite apart from food crises or other resources-disaster possibilities.
These comparative statistical judgments are, of course, essentially empirical and short-run. Whether they reflect stable long-run causal relations is at least highly problematic (though a number of specific case studies have suggested similar effects from commensurately large shifts in fertility). Much existing theory in this area—for example the classical "laws" of diminishing returns—has been irredeemably qualitative or deductive; practically none, including standard capital-output approaches, has been modeled sufficiently for statistical application and testing. Among the very limited numbers of statistically operational models that have been put forth, there are marked tendencies to utilize essentially the same—though untested—structural relations and to specify key parameters more nearly through crude empirical analogies with developed areas than by direct estimation. Such models probably overstate the influence of demographic factors affecting real savings and investment and understate such relatively "population independent" factors as exports and industrial development programs. Existing theory also understates the role of productive and distributive innovations and changing quality of work force, including the entrepreneurial sectors.
The magnitude of the macro gains to be expected in the low income countries from reduced fertility and slowing population growth over longer periods than a very few decades is probably impossible to determine reliably with present tools of analysis. What is needed are models that are predictively potent, yet can encompass the long lead times required before significant or even visible economic impacts can be traced to altered fertility trends. Assuming, for example, a sudden 25 percent decline in fertility, total personal consumption would not vary greatly in much less than 15 years (especially when the smaller consumption needs of children compared with those of adults are taken into account); educational and related collective consumption needs would stay largely unchanged for the better part of a decade, and
nftithpr lahnr-fnrrp «ivf nor rmmhprc of hnncphnlHc wnnlrl hp cnhctantiallw